GORAN CALIC
NSDA Lab
At Nuclear Systems Design and Analytics (NSDA) Lab we combine artificial intelligence, big data, and formal economic modelling to improve regulatory compliance and reduce cost and time overruns of nuclear power plant construction projects. We do this through two applications (NPP Construction App & NSDA AI) briefly described below.
NPP Construction App
The NPP Construction App (available here: https://npp.nsdalab.com/) is based on an AP-1000 construction costs and timelines white paper. This paper develops a scheduling and cost model for constructing a hypothetical AP-1000 plant. The model splits the build into stages, to illustrate the sensitivity of the build to factors like design changes or supply chain bottlenecks. The NPP Construction App is not intended to be representative of actual costs and timelines. It is a simple illustration of the formal model. Contact Goran for more information about the app and white paper.
NSDA AI
The NSDA AI is a user-friendly, chat-based AI platform designed to answer nuclear regulatory and design questions. It is trained on over 280 Million pages of regulatory and design documents. NSDA AI responses are intended to minimize hallucinations, which make traditional LLMs unreliable in highly-regulated, safety-critical environments. NSDA AI reduces hallucinations as a result of its training and demonstrates proof of its work by providing exact sources to design and regulatory documents it used to generate responses. Click on the slideshow to see example responses. Access to the NSDA AI is currently restricted. Contact Goran for access.
NSDA AI Screenshots
NSDA AI generates ideas using existing designs, submissions, and regulatory feedback.
The NSDA AI suggests 5 ideas, providing sources for each.
NSDA AI knows what worked, and what failed, in previous regulatory applications.
NSDA AI generates ideas using existing designs, submissions, and regulatory feedback.
